By the time the dot-com bubble burst, Sun had grown addicted to selling big, purple, multi-processor Sparc servers running Sun's proprietary Solaris OS. They were powerful. They were sexy. They were expensive. And, for a time, they were so successful that they allowed Sun to grow complacent and even to forget its own history.
Had Sun stuck to its roots, it might still enjoy the dominant share of the Unix market today. Instead it succumbed to its own early tactics. A new OS appeared -- Linux -- built from scratch but based on open standards. It could only do a portion of what Solaris could do, but it cost nothing and ran on commodity x86 hardware. Meanwhile, the x86 platform itself was evolving and becoming more powerful, to the point that clusters of commodity servers could begin to compete with the fastest supercomputers. While Sun gloated, the value proposition of its hardware was dwindling before its very eyes. The value of its shares wasn't far behind.
Thomas L. Friedman might describe it as another manifestation of the flattening of the world. Over the span of a few years, computing resources once reserved for the elites of Boston became available to their counterparts in Bangalore for next to nothing. How fitting, then, that the first stop on Sun's 2007 world conference tour should be India.
During his keynote in Hyderabad, Sun executive vice president of software Rich Green told the audience, "Open source is the vehicle by which you grow your volume." In other words, teach a man to fish and you build a market for bait. But how is a company as big as Sun to thrive in a market where even the fishing rods are seemingly commoditized?
The answer, of course, is to go back to its roots: open standards and open source wherever possible and raw innovation where it counts. Where it counts, in Sun's case, is the hardware. And what's driving it, as in 1982, is a crazy idea.
The current generation of computers does what it does well -- so much so that the full potential of modern processors is largely unrealized. That's one reason why virtualization is becoming so popular: Customers yearn to double up their server workloads and recapture those unused cycles. But Sun execs foresee a moment when the power of mainstream hardware will reach a critical mass that ushers in a new era of enterprise computing, one that will be every bit as significant as the shift toward commodity computing of the last two decades.
Sun CTO Greg Papadopoulos refers to this moment as the "Redshift." It corresponds to an explosion of what he calls "massive-scale systems." As enterprises begin to realize the full potential of modern processors, he says, supercomputing applications like complex computation, data warehousing, and grid computing will become commonplace. Internet-distributed compute farms will eventually grow large and powerful enough to operate on a global scale, serving applications that are only glimmers in the eyes of today's software engineers.